Why enterprise-architecture must be broader than just IT

Link: http://weblog.tetradian.com/2012/08/14/ea-must-be-broader-than-it/

From Tom Graves / Tetradian

Yes, I admit it, I made that classic mistake again: I allowed myself to get engaged in an EA discussion on LinkedIn…

This one started from the – to my mind – absurd and then insulting notion that EA is solely EITA (enterprise IT-architecture), and that therefore anyone who believes otherwise is a ‘dinosaur’. Not a good start – stating that the subset-scope is larger and more important than the superset – but that’s all too typical of the, uh, quality of argument we so often suffer on LinkedIn.

Anyway, the argument rambled on its its usual pointless way, with those who do work in whole-scope enterprise-architecture repeatedly struggling with the, uh, intelligence of the IT-centric folks. All rather sad and pointless, really.

And then, yes, it got worse. One of the regular choose-your-epithet types turned up, keen to… well, once again demonstrate why ‘choose-your-epithet’ applies. He was, as usual, arguing that the only thing that matters is IT, and that therefore all business revolves around IT, and therefore EA should concern itself solely with IT because there was nothing else that could matter to business. (Yep, you’ll note the nicely-circular argument there…)

I tried to illustrate that there might just be some business-concerns that might not solely be about IT, and that some industries cannot be centred solely around IT, by giving the real-world example of the logistics business, and the way in which mail-sorting and the like uses a mix of very complex electronic/mechanical systems and solely-manual sort-processes, with not much IT. He came straight back with this gem:

why not using emails instead ??? Is It an example from the 70s or something recent.

Yep: I blinked. What the…?

Heart sinking, I came back with a brief explanation that most mail-sorting is about physical stuff like letters, parcels and Christmas cards. It’s true there was a very small amount of mail that was IT-only – for example, some utilities-bills came in as a data-file and went out as printed/pre-addressed letters, but that was about it.

The reply, as you can no doubt guess, went straight back to IT, and then the same assertion that the only thing that matters is IT:

I just thought EAs understood that having an holistic approach is not realistic instead a lot of EA generally focus on essential capabilities design. Excepted for mail delivery companies I consider mailing a non-essential capability that I would not even dare analysing and design. And even UPS relied heavily on IT to implement their famous Hub and Spoke model (focused on transport more than on mail sorting).

… This group is just a virtual threat you cannot expect any deliverable from It but I do agree that some dinos need to adapt to the cyber-business.

Oh dear… the words “just doesn’t get it” obviously apply, but even that would be a wild understatement… Contrast that obsessive IT-centrism above with a trio of examples from people who do get that point – first from Carlos Manzanares:

Stating that EA = EITA makes the perception of EA as a technical practice rather than a ‘holistic enterprise’ practice even more entrenched, and in my opinion greatly devaluates its potential.

and Jeffrey Smith:

The whole purpose of doing EA to start with is to look at the entire business enterprise and how all the different pieces interact and can work together to improve overall efficiency. If you are not taking a holistic approach you can do this and you can not achieve the aims of doing EA to start with. What you end up doing is optimizing only one segment or domain of the architecture, often at the expense of the other domains.

and, more concisely, from Pearl Zhu:

the purpose of business is to serve people via people, IT is the means to the end, not the end

Yet IT-centrism remains so predominant in current ‘enterprise’-architecture that it seems the issue will always need to be addressed. For example, Alan Lloyd included this entirely-valid question:

is there a “substantial process” in any business today that isnt enabled by IT in some way ?

So what follows was my reply:

Not many – especially if we generalise it as ICT rather than solely [computer-based] IT. And therein lies a huge danger, as anyone involved in risk-management should recognise. The same applies to other core-utilities such as electricity, lighting and transport-infrastructure.

To my mind, it is absolutely essential that EAs should be engaged in risk-assessment and business-continuity/disaster-recovery [BCDR] planning, because – as Jeffrey and others say above – we’re one of the few disciplines that does (or should) take a fully holistic view of the overall enterprise.

Yet to do BCDR, the first requirement is to be able to imagine that the risk has eventuated, and that (in this case) the respective utility is not available. Only then can we ask meaningful questions about how long we can operate without that utility, what business-impact that would have, and priorities and practicalities for disaster-recovery.

If you have an ‘EA’ that assumes that IT always exists (somehow magically even where the utilities and infrastructure that underpin it don’t in fact exist…), then it becomes impossible to do that planning-work. Hence why I find the ‘IT is the centre of everything’ position so utterly infuriating: along with many other real-world problems, it guarantees risks to and loss of the organisation’s resilience.

So, once again, in the now-desperate hope that the IT-obsessed folk will finally get the point:

  • No-one in EA doubts (or should doubt) the usefulness and relevance of IT / ICT in the enterprise.
  • Any assessment of opportunity in the enterprise should and must include ICT, along with all other opportunity-domains such as business-model innovation, organisation-structure innovation, process-innovation and much, much else besides.
  • Any assessment of risk should and must include the possibility of the unavailability of of any or all of the core resources, capabilities, people and suchlike, including the unavailability of ICT.
  • Viable / valid assessment of context, opportunity and risk demands an awareness of the whole context, within and beyond the organisation and its business – which in effect mandates an holistic approach to all such concerns.
  • In an holistic approach to architecture (or anything else), everywhere and nowhere is ‘the centre’, all at the same time. Any attempt to place anything as ‘the centre’ around which everything else must revolve – such as the supposed centrality of ‘cyberbusiness’ and suchlike – will guarantee that the resultant architecture will fail. (To quote Jeffrey, “What you end up doing is optimizing only one segment or domain of the architecture, often at the expense of the other domains.”)

If any of these points are not obvious to you, and the validity and necessity of those points is not obvious to you, you should not attempt to do enterprise-architecture, because you will be putting your organisation at risk, almost certain to cause serious damage to the organisation and its business, especially over the longer-term.

If any of these points are not obvious to you, it doesn’t mean you’re wrong in itself: it just means that you should not attempt to do enterprise-architecture. So go do something else, such as ICT-oriented business-strategy and ICT-system design – you’ll probably be very good at it. But don’t call it ‘enterprise-architecture’, because doing so makes it much, much harder for those of us who do do enterprise-architecture to do our work. Okay?

Sigh…